Correlation, Regression, and Causation

In statistics, correlation is often misunderstood to imply causation. For example, today, the causes of childhood polio are well known; however, it was not that long ago that people, including some of those in the health professions, thought that eating ice cream might be a cause. Researchers noticed that in the summer two things happened: children ate ice cream and children developed polio. There was actually a fairly strong correlation between ice cream consumption and polio. As a result, desperate mothers were quick to ban the treat in the hopes of protecting their children. This is just one example of how correlation has been misunderstood to imply causation.

This week you will examine correlation and simple linear regression. Both of these statistical procedures can be used to measure the strength of an association between two variables, but neither one can be used to demonstrate causation.

for this Discussion, focus on the definitions of correlation and regression presented in your Learning Resources, how they are similar, how they are different, and for what each is used. Then, search for examples in the public health literature related to one public health issue of interest to you that illustrate the difference between correlation and regression.

post a brief summary of a public health example (from the literature). Then, compare correlation and regression using this example from public health to illustrate. Finally, respond to the statement, “Correlation is not causation,” provide an example from public health to illustrate.
Course Text: Essentials of Biostatistics in Public Health Lisa M. Sullivan


Correlation, Regression, and Causation